J. Fearnley vs T. Boyer — prediction

Fearnley's Elo advantage (1890 vs 1786, roughly a 104-point gap) and better ranking (159 vs 191) underpin the model's 65-35 lean, which sits close to the raw baseline split of 38-33. Boyer's ranking has also slid further and faster (-78 vs -34 for Fearnley), suggesting the gap in current standing is not just a snapshot but a trend.

Recent form reinforces the level gap: Fearnley is on a 6-match win streak (8-2 across his last 10) and has beaten Walton (Elo 1911) and Michelsen (Elo 1910), both rated above his own mark — a meaningful quality signal. Boyer, by contrast, comes in off a loss and a choppier 6-4 record over his last 10, with no listed quality wins.

The rest picture cuts against the favorite. Fearnley has played six matches in the last 14 days and enters with only two days since his last outing, a heavy recent workload that can erode serve execution and movement in a longer match. Boyer, meanwhile, hasn't played in 24 days and shows zero matches in the last two weeks.

That layoff for Boyer is a double-edged sword — it removes fatigue as a factor but introduces match-rust risk after three-plus weeks away. On balance, the concentrated volume of Fearnley's recent schedule is the more concrete red flag here, tempering the otherwise clear level and form edge in his favor.

The service numbers are nearly mirrored: Boyer wins 64% of service points to Fearnley's 63%, while Fearnley returns slightly better at 42% versus Boyer's 40%. Neither player profiles as a clear server-vs-returner mismatch in this pairing — the small return edge nudges toward Fearnley but is not, by itself, a deciding factor.

At odds of 1.47, the market implies about a 68% win probability for Fearnley, while the model's own estimate is a touch lower at 65%, producing a -5% expected value. This is a Challenger-tier Elo estimate, a softer, less-scrutinized market where any edge is unproven rather than confirmed.

Being the favorite here is not the same as holding value: the market is already pricing Fearnley slightly higher than the model itself does. The level, form, and h2h signals all point his way, but the rest asymmetry adds real uncertainty, and the pricing offers no discount to compensate for it.